Stupid question, what is a distilled model?

Yes-ish. The base is Draw Things and the relevant bits are https://github.com/drawthingsai/draw-things-community?tab=readme-ov-file#cuda-capable-linux that isn’t too difficult to setup. The app with the pretty interface is Apple only (the developer one day decided to cram the full 1.5 on his iPhone and that was the start of this. The app has feature parity between the iOS, iPad and Mac versions, the gRPC server is “just” the generation parts decoupled from the app) but there’s a Comfy plugin to use the server.

BTW on Apple’s hardware Comfy is poorly optimized, while Draw Things is optimized. The iPhone XR is the oldest hardware capable of on device generation, and (with the right settings) could do a SDXL 1024x1024 generation. 13 minutes mind you for 8 steps, but also 3gb of total system memory. On the other hand, the iPhone 17 Pro is a third of the speed of my RTX 3060. There’s also a friendly Discord, and the dev clearly enjoys adding support for new, cool models because he’s quick at it but doesn’t share roadmaps of any kind.

Yeah. I really, really like that thing.

Yeah, you can load Qwen Edit with Comfy UI. This tutorial for Comfy UI should help get you started.

Is there a way to run this in a Linux-based system?

Clearly the right choice is that I should totally have spent thousands of Eur paying someone to sculpt it for me me or thousands of hours learning how to do it myself

You can also find and use 2.1 but in my case I used https://makerworld.com/en/makerlab for the model conversion

Ah. Makes sense. Are you able to point me any resource? I found 2.0 on (what looks to me to be) their official site, but you had said 3.0.

The hunuan 3d is very complicated to run, so it's not currently available on the horde

Cool use of ai!

I don't see the Hunuan model on the AI Horse site. Is it on there and I'm just blind, or is it elsewhere?

Ok, this is freaking awesome actually!

I thought it was a candle, and now I need a candle of this.

Damn, that's really cool.

holy shit, that's really good

Someone uploaded a Tchnicolor LoRA.

I'll have to keep an eye on it, thanks!

It seems to work. I've seen some in-progress LoRA samples floating around.

I thought we couldn't do training until the base model was released?

Yeah, if you're using character LoRAs you're probably going to need to use regional prompting to keep the concepts from bleeding into each other.

A big note that it's nigh impossible to put two separate characters in one image, as the model has no idea who you are referencing. You can see this in all image models how they usually only have one central person, or two people who look oddly alike. Only way to get two or more characters would be with complex workflows to inpaint certain characters.

If you have or can create a LoRA trained on images of the character you're presenting, that may be helpful. Or if you have a checkpoint model trained on that character. Would be like having a character that the base model is trained on.

Right. That. Thanks.

You might be thinking of a LoRA. LoRA are adapters you use with larger models to help it generate whatever concept you want.

You would probably need to use something like Visual Novel Character Creation Suite to generate a character and a dataset for LoRA training, and then train a LoRA for each character. LoRA are adapters you plug into larger models to help it generate whatever concept you want, in this case it would be a character. With a well-trained LoRA you shouldn't have any problem with character consistency.

If you're trying to get started with generating locally, here's a tutorial for Comfy UI.

Eyeballing here. I'm a learner and have hardly used this.

Would a checkpoint model (right term?) achieve consistency, built from a specific set of pictures?